Identify trends among outlier genes

[sjspielman]

This issue is a placeholder to follow #129 and explore whether there are any biological patterns to outlying genes identified in bulk ~ pseudobulk models. Specific steps to be taken here will be fleshed out when we reach this stage but some initial ideas:

• test whether cell type marker genes (using our Panglao lists) are overrepresented in this set
  • we can use genes that show extremely high correspondence (perhaps taken as those points with lowest residuals in the model) as a background set
• hallmark GSEA is always a great low-hanging fruit